Simulation of self-driving agents negotiating speed changes approaching traffic-light-free intersections.

This proposal describes a research project about self-driving agents that negotiate with each other to improve traffic flow, safety and energy efficiency while approaching an intersection without traffic lights or a roundabout (NL: rotonde).

The idea is to simulate self-driving agents approaching an intersection, avoiding collisions by negotiating about the speed on time, maintaining a targeted average speed related to energy use and capabilities of their vehicles. When the agents come to an agreement, they will execute the appropriate actions stated in that agreement.

Every interaction starts by sending a message to the traffic controller acting as a deal maker between the agents by matching proposal and acceptance agreements.

It can happen that during the execution of the agreement, unexpected events occur. Agents are, in that case, requested to negotiate again, trying to mitigate the event.

An agent can send a proposal-agreement to the traffic controller, or an acceptance-agreement, taking into account the rules given by this traffic controller. These rules are based on the location, intended direction, and speed of the vehicles and are continuously monitored. Agents should send in their current speed and location, in addition to the agreement, to the traffic controller.

Minimum-viable: show a number configurations of three self-driven agents on a map approaching an interaction. Variables are speed, priority, issuing argeement bids to a traffic controller deal maker. Fully completed if the system can simulate eight vehicles under randomised conditions including migitating sudden unexpected events. Happy if I can see the vehicles moving errorsless while an dashboard is updated with traffic efficiency, flow and energy indicators.